Year: 2024
Type: dissertation
Abstract: This master's thesis studies potential futures for how artificial intelligence (AI) will affect knowledge work in the future. Specifically, the study experimented with the difference between two human AI financial analyst jobs where test users estimated the stock market effects of news articles. In one version users worked as responsible experts who have control over AI and can willingly decide when to use it. In the second version, users were supervising AI against false predictions and did not have control over seeing the AI prediction. The test scenarios were created based on a Future Wheels workshop following a methodology of Design for Uncertain Futures. The qualitative results of the test revealed a high tendency in people to make their own evaluations before consulting AI and feel more subjective effort when they are evaluating AI directly. A minority of participants found the scalability of teaching AI models to feel more significant than simply working as an expert. The results indicate that when human agency is important, AI collaboration should happen under worker control.
Source: Aaltodoc (Aalto University)
Author: Vahvelainen, Leevi
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7001600221